INDEX

Explanations

names of individuals or experts

oai_token-act-pair · gpt-3.5-turbo

references to individuals and their statements or roles

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 partying

-0.74

 murdering

-0.71

 matched

-0.65

 manipulating

-0.63

 discriminated

-0.62

 emulate

-0.62

 physically

-0.62

 obstruction

-0.62

 vested

-0.61

DragonMagazine

-0.61

POSITIVE LOGITS

 Shapiro

1.08

 Greenberg

1.07

 Beir

1.05

 Katz

1.03

 Feldman

1.03

 Levin

1.03

 Levine

1.03

 Cohen

1.03

 Kaplan

1.02

Lev

1.02

Activations Density 0.188%

names of individuals or experts

references to individuals and their statements or roles

No Comments

No Known Activations